AI Automation & Agent Workflow Builder based in Bangkok. I design and verify multi-agent systems on Google Cloud — ADK agents, Model Context Protocol servers, and telemetry — plus the n8n workflows that connect them to business tools.
No rounded-up claims. Each of these can be verified from a public repository, a recorded walkthrough, or a certificate registry.
Two tracks: the n8n automation systems that pay for themselves in saved hours, and the AI-assisted software products I direct from specification to release.
Automating inbound clinic calls to handle patient inquiries and book calendar appointments.
Outcome — Successfully connects call context to Google Calendar events and Sheets via n8n.
Automating industry news research and drafting persona-matched LinkedIn posts safely.
Outcome — Strips unwanted Markdown and routes clean data to Airtable and Supabase upon HITL approval.
Automating the searching of job boards, checking CV matches, and drafting cover letters.
Outcome — Successfully matches job descriptions and sends Telegram interactive approvals.
Handling incoming client requests efficiently by filtering out low-budget leads.
Outcome — Successfully filters leads via score routing and notifies via Telegram/Gmail.
Leads that arrive through a form go cold — nothing scores them, nothing follows up, and two separate CRMs drift out of step.
Outcome — Scores and tiers incoming leads, keeps Airtable and Google Sheets in sync, and runs the full 7-day follow-up sequence with Telegram alerts.
Eliminating manual spreadsheet logging using a natural language interface.
Outcome — Parses natural language into structured Notion/Sheets data with flow calculations.
Monitoring multiple AI news sources manually is time-consuming.
Outcome — Processes feeds and outputs Burmese text summaries.
People need a controlled way to create trackable short links and manage multiple public link-in-bio pages.
Outcome — Shipped against acceptance criteria I wrote and signed off — authenticated shortening, click analytics and QR sharing.
Enterprises require proven, secure agentic patterns to connect LLMs to unstructured knowledge bases, big data warehouses, and automated operational workflows without security risks.
Outcome — Three architectures I specified and verified on Google Cloud Run — grounded RAG with Vector Search, BigQuery SQL reasoning over an MCP server, and sandboxed Python execution. Selected as a featured Hack2Skill APAC submission.
Producing localized high-retention video ads and explainers traditionally requires expensive video crews, disjointed editing tools, and manual telemetry tracking.
Outcome — The 20 QA gates I defined decide what ships — the pipeline rejects its own output rather than publishing it. Built on Google Cloud Run; Google Cloud Agentic Cinema Hackathon (Devpost ClickHouse Partner Track) entry.
Flight cancellations and delays leave travelers stranded with complex rebooking procedures, confusing international transit visa requirements, and billions in unclaimed passenger rights compensation.
Outcome — 13 guardrailed skills under explicit capability limits — EU261 / UK261 / US DOT jurisdiction detection, visa-aware rebooking against a curated visa table of 14 passports, and regulation-cited appeal letters. Alibaba Cloud x Atlas Agentic AI Hackathon entry.
The same process whether it is a two-week n8n build or a multi-agent system on Cloud Run.
I map the current manual work, find where the hours actually go, and say plainly which parts are worth automating and which are not.
I choose the workflow, the APIs and the models — n8n, Vertex AI, Claude, Qdrant — and write down the failure modes before any code is written.
I direct AI coding agents — Codex (my main one), Antigravity, Claude Code, ZCode, OpenCode — to implement it, then judge the result against QA gates I define, deploy, and hand over the documentation you need to own it.
Client services, project coordination, hospitality — then a deliberate move into AI automation. The operations background is why I design for handover, not for demos.
I moved from Client Services and Project Coordination into AI Automation.
I run every build as a project with a scope, a spec and a QA gate. That is the part of my coordination background that transfers directly: automation fails on ambiguity far more often than it fails on code.
I design, implement, test, and troubleshoot n8n workflows and agent systems with API, webhook, LLM, RAG, and human-in-the-loop integrations.
Every credential below lists its issuer and issue date. 11 of them link straight to the issuer's public verification page — check them yourself.
Anthropic Skilljar · August 2026
Anthropic Skilljar · August 2026
Anthropic Skilljar · August 2026
Anthropic Skilljar · August 2026
Anthropic Education · August 2026
Anthropic Skilljar · August 2026
Google / Coursera · March 17, 2026
UiPath / Coursera · April 3, 2026
Vanderbilt University / Coursera · September 28, 2025
DeepLearning.AI / Coursera · September 23, 2025
University of Michigan / Coursera · November 27, 2025
Anthropic Education · completion verified
M.H.T.I
August 2022 to November 2023
East Yangon University — degree not completed
February 2017 to March 2019
A selection someone else made, and a community I build in the open with. Both are public, so check them rather than take my word for it.
Selected as a featured submission for the Accelerate AI with Cloud Run track at the Hack2Skill APAC GenAI Academy. The choice was theirs, not mine — the architectures themselves are in Selected Work.
Contributing to an open-source developer community — building on shared material and reviewing each other’s work in the open. The closest thing to peer review I have, and it is public.
Source